App Disaggregation’s Physical Limit: 30 apps per device

eBay Marketplaces: 3 apps (Core, Fashion, Motors)
Facebook: 5 apps (Core, Messenger, Whatsapp, Instagram, Paper)
LinkedIn: 3 apps (Core, Contacts, Recruiter, and until a few days ago, CardMunch)
Google: 36 apps (ridiculous)

Across mobile, the companies most serious about mobile are disaggregating their features sets and building new, more focused, apps. There are compelling reasons to split up your apps:

  • Apps get one chance to get key permissions from the user — push and geo, in particular. With multiple apps, you can ask multiple times, and users might have a clearer understanding of why the permissions are necessary. And if one of a company’s apps annoys users, the users are less likely to totally cut out that company (by revoking permissions in all apps).
  • Smartphone screens are small — even on the largest Android devices. Focused apps can have simpler user interfaces, without hidden features. If a button isn’t visible, most users will never find it. This is why Facebook removed their slide out navigation panel and why Venmo’s right nav surprised me. RedLaser has a left navigation, but we find that users are orders of magnitude less likely to use any features there.
  • Updates take time on iOS. If you’re testing new features which might anger customers or be unstable, it’s better to test in less important apps and eventually roll the best and most stable features into the core experience.

However, most users don’t have many apps on their devices: Surveys suggest users have only gone from 31 to 33 apps on their devices over the past five years. (See my data here.) Moreover, most of these apps are likely games.

Average Apps per Device

If users don’t have more apps, who is using these more focused apps? A few hypotheses:

  • Power-users use anything other than the core apps (and maybe Whatsapp). App manufacturers can create power tools for some users without distracting most of their user base.
  • Apps from powerhouse developers are beating out smaller companies. eBay/Facebook/Google/etc. have a larger share of the install base than they did 4-5 years ago.
  • None of the ancillary apps have significant install bases. They’re purely used for experimentation, and all the successful features will be rolled into the core apps.
  • The cohort of users with smartphones in 2010 might have many more apps installed in 2014 — but the average is weighted down by new late-adopting smartphone users.

I suspect all these explanations hold a bit of truth. It’ll be an interesting place to watch — particularly because discoverability of installed apps becomes a huge issue once users have 100+ apps. I’m already at the point where any apps off my home screen are accessed via Apple’s built in search feature or Siri.

Why freemium’s only going to get more common

Localytics, via Techcrunch
Evernote blog, via Markitecture

These two charts, taken together, are what have me convinced that freemium is going to be the dominant business model for any serious application in the future.

What these charts tell me:

  1. Mobile applications either win or lose; users try lots of apps, decide which they like, and then use those few heavily
  2. Users are willing to pay for apps they use heavily

The key, then, is how to convert users into heavy users, and thus convince them to pay. Evernote, Dropbox, Pandora, and others find the freemium model well-suited to this–get lots of free-users, offer a compelling product, and convine the heaviest users to shoulder the majority of the cost.

There are other options. Microsoft and Adobe are able to jump immediately into a relationship via their reputation and business lock-in. But I suspect the days of such uniformity are drawing to an end. Google Docs, Evernote, and a whole suite of applications threaten the Office lock-in. Adobe and PDFs, another must-have software set, is also on the decline.

I can imagine the freemium model expanding to other industries. Amazon’s announcement of free book rentals is an interesting stab at a physical-world freemium program–users still need to buy (subsidized) Kindles, but it seems like their goal is to lower the trial costs as much as possible, and encourage the heaviest users to come in with gusto.

There’s real money to be made in figuring out how to apply this pay-gradation to other industries. I could imagine it would be effective in video, in particular.